2000 Florida Anthropological Society Inc.
The Florida Anthropological Society Inc. holds
source text of the Florida Anthropologist
considered the copyright holder for the text
all rights to the
and shall be
and images of
The Florida Anthropological Society has made this publication
available to the University of Florida, for purposes of
digitization and Internet distribution.
The Florida Anthropological Society reserves all rights to this
publication. All uses, excluding those made under "fair use"
provisions of U.S. Code, Title 17, Section 107 are restricted.
Contact the Florida Anthropological Society for additional
information and permissions.
A HRO OLOG IST
PUBLISHED BY THE
FLORIDA ANTHROPOLOGICAL SOCIETY, NC.
September 197 6
THE FLORIDA ANTHROPOLOGIST is published quarterly in March, June
September, and December by the Florida Anthropological Society, Inc., c/o
Room 102, Florida State Museum, University of Florida, Gainesville, FL 32611.
Subscription is by membership in the Society for individuals or institutions in-
terested in the aims of the Society. Annual dues are $6. 00; student members
t4. 00. Requests for memberships and general inquiries should be addressed to
the secretary; subscriptions, dues, changes of address and orders for back issues
to the treasurer; manuscripts for publication to the editor; and newsletter
items to the president. Second class postage paid at Gainesville, Florida.
XXIX No. 3
Florida Seminole Silver Work by Byron A. Johnson . .. .93
An Archaeological Survey of Upland Locales in Gadsden and
Liberty Counties, Florida, by George W. Percy and M.
Katherine Jones . . . . . 105
Spanish Trade Pipes from Marion County, Florida, by Wilfred
T. Neill and George R. Ferguson .. . . 126
The Seminole Pumpkin by WilfredT. Neill . .. . 129
OFFICERS OF THE SOCIETY
President Wilma B. Williams
2511 McKinley St., Hollywood, FL 33020
1st Vice President Raymond Williams
Dept. of Anthropology, University of
South Florida, Tampa, FL 33620
2nd Vice President Jerald T. Milanich
111 SW 23rd Terrace, Gainesville,
Secretary George W. Percy
Div. of Archives, History, and
Records Management, 401 East Gaines
Street, Tallahassee, FL 32304
Treasurer Norcott Henriquez
1510 Dewey St., Hollywood, FL 33020
Directors at Large
Three years: Robert E. Johnson
4250 Melrose Avenue
Jacksonville, FL 32210
Two years: Ray C. Robinson
1020 4th Street North
St. Petersburg, FL 33701
One year: Wesley Coleman
10 NW 124 Avenue
Miami, FL 33126
Editor Ripley P. Bullen
102 Florida State Museum,
University of Florida
Gainesville, FL 3261
FLORIDA SEMINOLE SILVER WORK
Byron A. Johnson
The period of the 17 through the 19th centuries was a time of European
encroachment onto Indian lands along the Eastern Seaboard and the Gulf Coast
of North America. European contact with Indian tribes had the ultimate effect
of disintegrating the tribal structures and causing forced or voluntary migration.
The Creek and numerous other fragmented elements of the Muskhogean
speaking tribes were faced with two possible alternatives: migrate westward
into an unfamiliar ecological area, or move further south into the desolate, but
familiar, interior of Florida. Slowly, in small groups, a large number of Mus-
khogean and Hichiti speaking Indians moved south to Florida where they were
joined by other Georgia and Alabama Indians to form what we refer to as the
Seminoles. This new tribal entity reached its zenith during the 18th century.
Numerous writers (Mahon 1967, Mac Cauley 1887, Nash 1931, etc.) have
commented at length on the arts, crafts, and costumes of the Seminole. How-
ever, one area of their material culture has received little attention, that of
the Seminole craft of silver work.. The craft requires closer inspection because
of the delicacy and state of the art which it achieved, with little or no European
influence on technique or style.
The origin of silver work among the Seminole is unknown, but it is likely
the result of three separate influences: (1) early pre-Columbian work with nat-
ive copper, (2) 18th century Calusa work with Spanish coin silver, and (3) Euro-
pean trade and military influence. Of course, the Seminole were familiar with
silver ornaments before they came to Florida.
Native copper is found in deposits around the Great Lakes and has been
known since pre-Columbian times. Goggin (1940:25-26) noted the high inci-
dence of Indian experimentation with native copper which resulted in the pro-
duction of knife blades, spear points, vessels, and breast pieces.
After European contact the Calusa Indians began working with silver from
Spanish coins. Whether they made the transition from copper work to silver
work by themselves or were taught silver work by the Spanish remains a mys-
tery. However, the manufacture of silver by the Calusa in the form of pendants,
gorgets, and beads became a highly developed art. The Seminole may have had
some knowledge of this (Fig. 4).
Finally, the European and American military forces and the Indian traders
Florida Anthropologist, vol. 29, no,3, September 1976
seem to have had a formative effect on Seminole silver work. It was a standard
practice in the 18th century for colonial powers to seal a military pact with an
Indian nation by giving the major chief or chiefs one or more silver military
gorgets (Woodward 1926:232-249). These gorgets became so popular among the
Seminole that they began to fashion imitation military style gorgets of their own
from beaten coins. Osceola and other prominent chiefs are usually depicted
wearing sets of silver military gorgets. (Fig. 1).
Indian traders of the 18th century either gave or sold large quantities of
silver articles to the Seminoles, thereby giving them models for future elabora-
tion. One 18th century trader remarked at length about giving Seminole tribes-
men "...silver armplates, wrist plates, gorgets, and earbobs..." Later sil-
ver work displays a combination of European influence on form and Seminole de-
velopment of specific design practices.
The Manufacture of Silver Work
The social organization behind the manufacture of silver work in Seminole
society was not elaborate, nor were there any restrictions placed upon the sil-
versmith. No information exists concerning how persons obtained tutelage in
the craft, but Goggin notes that some aspects of the craft were passed down from
father to son and that silver work was practiced by men only. Certain smiths,
such as Billy Mathla, obtained fame within the tribe and passed their knowledge
and skill down to their sons. When the craft died out in the 1930's the tradition
of father to son transmission of knowledge was apparently broken.
Silver work was produced on a contract basis when required by a member
of the tribe. The amount or form of payment for the work is unknown, but it
is likely that goods were bartered for the finished spangles instead of currency
and coin which was reserved for the trading post.
The source for silver used in the pieces of silver work was coinage of low
denomination. Indians at trading posts, such as Storter's at Everglade, "...
would refuse paper money, since it was subject to rotting or being mutilated in
the camps. Gold and silver coins were relatively indestructible, and they could
be strung together as ornamental necklaces or bracelets (Kersey 1973:262). The
denominations known to have been used are: United States half dimes, dimes,
quarters, and half dollars; British shillings and sixpence; and Spanish coins of
unknown denominations. The favorite type of coin seems to have been the United
States dime because of the high silver content, availability, and inexpensiveness.
Skinner's description of the working of silver is the best contemporary ac-
count and is in part repeated here: (Skinner 1913:74-76)
"To make a spangle, a coin is heated in a small fire; it is then removed
Fig. 1. Osceola with silver military gorgets
with a pair of pinchers and hammered out with an ordinary commercial ham-
mer. The poll of an axe driven into a log serves the purpose of an anvil. The
process of alternate heating and pounding is repeated again and again until the
coin has been flattened out considerably and the design effaced. One smith ob-
served at work greased the coin from time to time as he heated it. After it has
been heated and hammered to the satisfaction of the smith, the spangle is pared
down with a butcher-knife or a razor blade until it has been reduced to the de-
sired degree of thinness.
"In this state the blank form is sometimes decorated with a design incised
with a file or a knife blade. Any irregularities are filed off and the trinket is
polished on a whetstone. Sometimes the designs are cut out with a cold-chisel
and finished with a knife. Holes for sewing the spangle to a garment are made
by driving a nail through the metal and smoothing the edges with a knife.
"This process of silverworking was observed on two occasions, and there
was but little difference in the tools or in the manipulation of the smiths. Antler
prongs are used as punches to make raised lines and bosses, and the only other
tool which was seen or collected besides those described, was a crude blow-pipe
used in the manufacture of the plain finger-rings which are much worn by the
In order to determine the control over uniformity which the Seminole sil-
versmiths were able to attain a group of circular bodice-pieces and geometrical
pendants were selected for study from the extensive collection of the Historical
Association of Southern Florida located in Miami. The measurements recorded
included diameter, thickness, weight, and percentage of paring (Fig. 3.).
Fig. 3 Bodice-piece and pendent measurements. [Here Johnson presents
a list of 54 items including 3 half dimes, 50 dimes, and 2 quarters. The last
had diameters of 44.3 and 49 mm, a thickness of .4 mm, weights of 5.27 and
4.97 gm, and paring percentages of 15.5 and 20.35 respectively. The half dimes
were similarly 22.7,22.6, and 23.2 mm in diameters; had a thickness of .4,
.3, and .3 mm; weighted .92, .93, and .84 gm; and paring percentages of 11.8,
25, and 33.87 respectively. The dimes varied in diameter from 22.6 to 33.7 mm
with most between 31 and 33 mm with exceptional example listed as having a di-
ameter of 49.2 mm (it must be a quarter). Thickness of dimes ranged from .2
to .7 mm's with a preference for .4 mm except for the freak at 1.4 mm. Weight
varied from 1.77 to 2.37 gm except for the large one which had a weight of 4.75
gm (similar to that of quarters). Paring percentages went from 5.2 to 44.8
with most between 10 and 20.]
The control of the artisan over his medium was found to be excellent. De-
viation in the diameter of bodice-pieces (made from the same denomination of
coinage) was found to be within 5 mm in every case. Thickness of the bodice-
pieces, measured at the rim, was found to be no less than .3 mm and no great-
er than .6 mm. After comparing the weights of the finished pieces to the ori-
ginal weights of the dimes and quarters which the pieces were made from it was
found that approximately 200/o of the silver had been pared away. Approxinately
300/o of the silver had been pared away from the geometric pendants in order
to form the cold-chiseled designs mentioned by Skinner.
Types and Styles
In the century and a half that the Seminole produced silver jewelry they
explored an amazing number of styles and types which went in and out of fashion
at a tempo that rivals modern trends in wearing apparel. The following is a com-
pilation of what is known about the various types and styles of Seminole silver
Pendants: Pendants were both circular and rectangular in shape and could be
either solid or cold-chiseled into many varieties of designs. The women hung
the pendants from beaded necklaces in combination with perforated coins or by
themselves. Men wore the pendants on watch chains stretched across their shirt
shirt in the fashion of the modern day charm bracelet, or used them to fringe
sashes. Late in the 1920's and 1930's the women attached the individual pen-
dants to their blouses in the form of broaches. Known cold-chiseled designs
are both geometric and realistic and include: crosses, stars, hearts, arrows,
swasticas, sailing canoes, anchors, and initials (Figs. 2-3). Goggin confused
the pendants with bodice-pieces; the pendants are hung free or pinned while the
bodice pieces are sown on the front of the bodice semi-permanently.
Earrings: Earrings were worn predominantly by women in later times. Most
of them were made into geometric ovoid shapes, but a few were cold-cheseled
into designs similar to pendants. The usual size of the earrings is under 1 inch,
but unless a piece is documented it is difficult to separate an earring from a pen-
Turban Bands: Turban bands were in fashion in the 1830's and were present on
virtually every picture of a Seminole warrior painted during that period. They
were 1-1.5 inches wide and held together by a leather thong. Some of the sur-
viving specimens are decorated with scalloped edges, lines, and dots. Goggin
reports that they were very rare around 1940 (Goggin 1940:28). Since many
Creek warriors wore turban or hat bands it is possible that the style may have
been brought to the Seminole by 19th century Creek migrations.
Bodice-Pieces: Bodice-pieces were strictly women's jewlery in the form of con-
vex discs of silver with incised lines and/or scalloped edges. They range in
size from 22.6 to 44 millimeters and there are 5 varieties: (1) an incised line
and outer punctations (Fig. 5, B), (2) an incised line and scalloped edges (Fig.
Fig. 2. Seminole pendant designs
r~^ I M""1
------ V 4 vvV
1! 4~ rr \t -ww
^^j \\ y/J
Fig. 3. More cut work designs
I!111 !l l '2.l T! t 1 tll l7li l l i If i
Fig. 4. Calusa silver pendants
..... ... ,HA S .
Fig. 5. A, geometric pendants; B, punctated bodice pieces; C, scalloped
Fig. 6. Inverted "V" stamped bodice pieces
&.ss4r'r V.. rl .t -t' '
Fig, 7. A, large Type 5 bodice pieces; B, Type 5 bodice pieces; C, bodice
pieces with more than two holes
5, C), (3) a stamped, inverted "V" shaped impression (Fig. 6), (4) a raised
center, and (5) a single stamped round line. The pieces were sown onto the
bodice by two holes, one small and one large, driven through the center of the
piece. In a few of the pieces from the collection of the Historical Association
of Southern Florida there are more than two holes for an unknown reason. Wo-
men were wearing the bodice-pieces as far back as the 1880's and the design
appears to have changed little. The number worn seems to have been highly
variable; the Historical Associations collection of pictures taken by Claude C.
Matlack during the 1920's and 1930's shows individuals wearing as few as four
and as many as twenty-five bodice-pieces.
Crescent Gorgets: Military style crescent gorgets are among the oldest type of
Seminole silver work (Fig. 1). They seem to have gone out of favor after the
Seminole wars and were exceedingly rare by the year 1900 (Goggin 1940:29-30).
Worn in sets of three they were decorated by a single line and bossed ends.
Bracelets: Bracelets were small silver bands 1-1.5 inches wide worn on the
wrist or upper arm. They were either soldered together or fastened together
with leather thongs. They are reported rare in the 1940's (Goggin 1940:30).
Rings: Seminole rings were simple soldered bands from .124 to .187 inches
wide and having a tablet extending from knuckle to knuckle (MacCauley 1887:490).
Baretta: It is possible that a type of silver baretta could have been worn after
1900. The only clue to this is a picture taken by Matlack about 1930 of a group
of girls visiting Watson's Island in Miami. No examples are known in any col-
lection and it is possible that the barettas were commercially manufactured.
Wear of Silver Work
When considering the role of an object of material culture in society quest-
ions arise as to who could wear certain items of apparel and what they signified.
In Seminole society silver work was worn by both young and old alike. Both the
young and the old wore large quantities of silver and it does not seem to have
represented wealth or marital status.
The most important statement that can be made with regards to the import
of the silver is that it was valued highly. The Matlack photographs picture some
individuals, such as the matron Annie Tommie, wearing much silver during the
Green Corn Dance but only a pendant at other times. Silver evidently was worn
only during important occasions and events.
End of Silver Work
The end of the craft of silver work in Seminole society took place during
Fig. 8. Seminole coin necklace
the 1930's. Though it cannot be definitely confirmed, the effects of the depress-
ion may have caused a shortage of silver which prevented the elaboration and
continuation of the craft.
There do not appear to be any silversmiths living in the Seminole tribe
or the closely allied Miccosuki tribe. No one at either tribal headquarters knew
of anyone who had practiced the art in recent times.
The silver work seems to have gone in several directions: (1) it was taken
to the trading posts in times of financial difficulty and later sold by the trader,
( 2) it was buried with the older persons as they died, ( 3) it was sold to indi-
viduals and is in either private collections or museums, (4) it has been kept
on the reservations in small quantities as heirlooms.
The Seminoles have had a history of rapid fashion changes which have al-
tered the costume of the tribe in a short period of time. Silver work went out
of vogue and nothing has, as yet, replaced it. It is possible that the Seminole
silversmiths advanced the art of silversmithing by the process of pounding coins
under heat to the highest level ever achieved.
The author would like to express appreciation to the Historical Association
of Southern Florida, Inc. for access to its large collection of Seminole silver
work and extensive library. Special appreciation is expressed to Ms. Patsy
West, Curator, for help in locating indispensable texts and establishing contact
with the Seminole and Miccosuki tribes. Appreciation is expressed to Mrs.
Clara Lee Tanner, University of Arizona, for comparative information on the
Navajo art of silversmithing.
Goggin, John M.
1940 Silverwork of the Florida Seminole. El Palacio, vol. 40, No. 2,
Kersey, Harry A., Jr.
1973 Pelts, Plumes, and Hides: White Traders Among the Seminole
Indians, 1890-1930. The Florida Historical Quarterly, vol. 51,
no. 3, pp. 250-66. Gainesville.
1887 The Seminole Indians of Florida. 5th Annual Report, Bureau of
American Ethnology. Washington.
Mahon, John K.
1967 History of the Second Seminole War, 1835-1842. University of
Florida Press. Gainesville.
1931 Survey of the Seminole Indians of Florida. U. S. Government Print-
ing Office. Washington.
1913 Notes on the Florida Seminole. American Anthropologist, vol. 15,
no. 1. Washington, D. C.
1926 Indian Use of the Silver Gorget. Indian Notes, Museum of the
American Indian, Heye Foundation, vol. 3, no. 4, New York.
May 28, 1975
AN ARCHAEOLOGICAL SURVEY OF UPLAND LOCALES
IN GADSDEN AND LIBERTY COUNTIES, FLORIDA
George W. Percy and M. Katherine Jones
The purpose of this paper is to summarize the results of archaeo-
logical site survey in adjacent portions of northwestern Liberty and south-
western Gadsden counties. Both counties are located in the center of the
Florida Panhandle and are bordered on the west by the Apalachicola River,
which flows from north to south and empties into the Gulf of Mexico. The
Apalachicola is formed by the junction of the Flint and Chattahoochee and
is the largest stream in Northwest Florida.
The survey work reported here was part of a broader program of
archaeological research in the upper Apalachicola River basin which began
in 1971 under the direction of the senior author and the sponsorship of the
Department of Anthropology at Florida State University. This program has
been briefly described in several earlier papers by Percy (1971a; 1971b;
1972a; 197b; 1972c) and, also, Jones (1974); a more complete description of
its rationale and aims is currently in preparation. The general aim was to
learn about prehistoric cultural development in an inland region of North-
west Florida. Previously, most archaeological research had been restricted
to immediate coastal locales with the result that information on the pre-
historic cultures of the Panhandle was heavily biased in favor of a very
limited geographic area. Most of the interior of the Panhandle, comprising
a much larger geographic area and a diversity of environmental zones, all
extremely different from those of the coastline, was (and still is) practically
The upper Apalachicola basin project area (Fig. 1) takes in roughly the
northern half of the Apalachicola River itself and the areas on either side
which it drains. The northern limit is purely arbitrary and is set at the
boundary of Florida with Georgia and Alabama; thus, the west bank of the
lower Chattahoochee River is included in the study area, along with both
banks of the upper Apalachicola. The southern limit is in the vicinity of
Blountstown and Bristol, in Calhoun and Liberty counties, respectively,
where the transition from the Northern Highlands physiographic zone to the
Coastal Lowlands zone begins (Puri and Vernon 1964:10-15 and Fig. 5); the
general character of the environment starts to change rather significantly.
The western boundary is the watershed between the Apalachicola and Chocta-
whatchee drainages; the eastern boundary is the watershed between the Apa-
lachicola and Ochlocknee drainages.
Within these general limits, there are a number of distinct environ-
Florida Anthropologist, vol. 29, no. 3, September 1976
Fig. 1. Map of upper Apalachicola River region showing project area.
PERCY AND JONES
mental zones including the Apalachicola River itself; its tributary streams;
the banks and levees of the Apalachicola and those of its tributaries; the
rim swamps in the low area on the east side of the Apalachicola, between
the east levee and the bluffs of the bordering uplands; the bluffs themselves
and their talus slopes; and the hilly upland area on top of the bluffs. To
the west of the Apalachicola, in sharp contrast to the upland zone on the
east, is a broad expanse of relatively flat low-lying land known as the Mar-
ianna Lowlands. Here, in addition to the sluggish streams which presently
exist, there are swamp areas representing former channel positions of the
Apalachicola and tributary creeks. In the northern and southwestern parts
of Jackson County, there are small lakes and ponds which have formed in
sinkholes; in the northeastern portion there are also numerous small caves
and several important areas of limestone outcropping which contain flint
deposits. The present environment of the entire upper basin is basically
similar to what it was in, at least, later prehistoric times, and the differ-
ent zones would have offered a variety of possibilities for use by prehistoric
To date, archaeological investigations have concentrated in the upland
zone east of the river in the vicinity of Torreya State Park, which is in
the far northwestern corner of Liberty County. The terrain is typical of
the Northern Highlands. It is very hilly with sharp changes in elevation
and few sizeable areas of level land. The continuity of the land surface is
broken by many creek drainages with steep sides. The creeks are mostly
small, shallow, rapidly running streams which flow in a westerly or north-
westerly direction and ultimately empty into the Apalachicola. Throughout
the zone, the soils on hilltops and ridges are predominantly sandy; they
are very rapidly drained and are low in plant nutrients. Soils in the narrow
creek bottoms are more silty and are more favorable for plant growth. The
dominant vegetation over the southern half of the upland zone is an associ-
ation of longleaf pine, xerophytic (scrub) oaks, and wire grass, although
in creek bottoms a variety of hardwoods occurs. In the northern half of
the upland zone, forests of mixed hardwoods and pines are more charac-
teristic; this is also the case in the Marianna Lowlands.
The general purpose of the investigations in the uplands has been to
test the hypothesis that this zone was a frontier or marginal area for late
prehistoric agricultural groups inhabiting the upper Apalachicola basin; the
focus is particularly on Weeden Island and Fort Walton groups, since it is
assumed that agriculture was not conducted on a large scale and did not
play a crucial role in prehistoric Indian subsistence in Northwest Florida
until Weeden Island times (Percy and Brose 1974:19-22). Compared to other
parts of the project area--for example, the Apalachicola River valley--the
resource potential of the uplands is low. Soil quality is poor, except for
very small areas in creek bottoms; the vegetation is dominated by species
SURVEY OF UPLAND LOCALES
which do not produce edible nuts and seeds -- or, at least, these are not
available in anywhere near the large quantities that can be obtained in the
lowlands to the west; the density of terrestrial fauna is low; and the small
streams do not offer much variety or quantity of fish or shellfish. In other
words, the uplands are a comparatively poor place to live, especially for
peoples requiring large tracts of naturally fertile land. In Weeden Island and
Fort Walton times, small groups might have utilized the area for brief hunt-
ing forays or woodgathering trips, but their occupation would always be
transitory and short-lived and, perhaps, irregular. Archaeologically, this
type of usage should be represented by small camping stations, marked by
very light accumulations of refuse, little variety in cultural items, a low
incidence or absence of structural features, and very simple site plans,
displaying little internal organization. Carefully planned, year-round or
long-term seasonal base camps and other signs of more stable occupations,
such as mounds, should be generally absent. Instead, these should be found
in more productive resource zones, as along the levees of the Apalachicola
or Chattahoochee rivers, where large tracts of fertile land occur, and where
large quantities of wild plant and animal foods can be easily obtained from
nearby hardwood swamp forest.
This hypothesis, which is outlined only in a very brief form above, is
linked to a broader hypothesis concerning subsistence practices of late pre-
historic, especially Weeden Island, peoples in the upper Apalachicola basin.
This will be presented in detail in a forthcoming final report on the results
of the 1971 field season.
The investigations in the upland area east of the Apalachicola River
have so far involved two seasons of excavation (1971 and 1972) at the Torreya
site (8Li8) a Weeden Island village with associated burial mound (8Li5) in
Torreya State Park, and intensive survey of several creek drainages in the
vicinity of the park. The most important of these include Rock Creek and
the Upper Sweetwater and Big Sweetwater tributary systems of Sweetwater
Creek in Liberty County. In addition, smaller portions of the Crooked Creek,
Short Creek, Long Branch, Little Sweetwater Creek, and Kelley Branch
drainages have been surveyed, as have parts of four small unnamed creeks.
The locations of all of these streams and the areas along them that have
been visited are shown on the scaled map in Figure 2. The inset of Little
Sweetwater Creek and Kelley Branch is located about 11/4 miles below the
southern limit of the area shown in the main part of the figure. The total
area covered, up to the present time, is approximately 14 1/4 square miles,
nearly all of which, about 13 1/4 square miles, is in Liberty County, south
of Torreya State Park. Only about one square mile is in Gadsden County,
and much of this was covered in a separate survey of the Interstate 10
right-of-way by B. Calvin Jones of the Division of Archives, History and
Records Management. B.C. Jones, who did his work in 1972 and has yet to
PERCY AND JONES
FIG. 2 UPLAND SITES IN S.W GADSDEN
N.W LIBERTY COUNTIES
A BURIAL MOUNDS
S* HABITATION SITES
CONTOUR INTERVAL 50FT
Th7 5 p a. ad. ord forom ,e U s Geoogcal
75 mime serles
SURVEY OF UPLAND LOCALES
publish the results, has kindly allowed the authors to present some of his
data in this paper. He is responsible for locating the following sites on the
map: 8Gd12, 8Gdl3, 8Gdl4, 8Gdl5, 8Gdl6, 8Gdl7, and 8Gd89.
Of the various creek drainages, it is the Sweetwater-Upper Sweetwater-
Big Sweetwater system that has been the most extensively investigated. It
accounts for approximately nine of the total of 14 1/4 square miles. This
work was done by M.K. Jones, and the results have been presented in de-
tail elsewhere (Jones 1974).
All of the surveyed areas shown in Figure 2 have been covered on
foot, and the coverage has been complete, or as nearly complete as poss-
ible in the ardous, heavily wooded terrain of the uplands. This figure con-
veys to some degree the broken character of the land surface. The dense-
ness of the forest cover is shown in Figure 3, which is an aerial photo-
graph of the same area shown in Figure 2, excluding the portion in the in-
set. The two figures can be compared by matching the bends in the Apala-
chicola River. Also, Upper Sweetwater Creek, Rock Creek, Short Creek,
and Crooked Creek all show well in the aerial; Flat Creek is just above the
top of the photo. The bluff line along the east side of the Apalachicola also
stands out well. A total of 58 prehistoric Indian sites has been found, and
the authors think it unlikely that additional ones will be discovered in the
already surveyed tracts. This is not only because of the care with which
the original survey passes were made, but also because in the few instances
where tracts have been rechecked, as in Torreya State Park, the return
passes have not resulted in the discovery of new sites.
In general, the collection of artifacts from individual sites has been
minimal and confined to surface materials. The method of collection has
been a simple grab sample approach, in which a selection of artifacts is
gathered unsystematically from various parts of the site surface, and the
artifacts are lumped together as one provenience unit. Only the larger or
more obvious artifacts are chosen; thus, for example, decorated pottery
sherds tend to be favored over plain ones. There are obviously a number of
disadvantages to this approach (Ragir 1967: 182-84 and 191-94), one import-
ant one being that artifact distribution patterns, potentially reflecting
the occurrence and arrangement of activity areas within a site, are not re-
covered. There is also no way of appraising how well a collection may rep-
resent the full range and frequency distribution of classes of artifacts at
a site; (assuming, in the first place, that the populations of surface arti-
facts at these sites are sufficiently similar to the subsurface populations,
so that they are capable of yielding representative samples of the contents
of the sites).
The emphasis, to date, has been on site location and preliminary
PERCY AND JONES
State of Florida
SURVEY OF UPLAND LOCALES
characterization of artifact density and variety. Much future work will be re-
quired to develop reliable pictures of the sizes and contents of the various
sites. Present size estimates are based on observations of the spatial ex-
tent of surface materials which are always of low density, and the estimates
are quite rough and possibly very inaccurate. Cultural classifications and
information on the ranges of artifact types are also based mostly on small
surface collections, although M. K. Jones conducted limited test excavations
at six sites in the Upper Sweetwater Creek drainage (Jones 1974:87-112).
These include 8Lil5, 8Lil7, 8Li26, 8Li30, 8Li32, and 8Li34. Limited test-
ing has also been done at 8GdlZ (personal communication, B. Jones, Feb-
ruary 5, 1975), although no results have been published. The testing was
sponsored by the Division and was actually done by David S. Brose of Case
Western Reserve University. Extensive excavations have been done at the
Torreya site (8Li8) by Percy (1971a; 1971b; 1972a; 197b; 1972c; 1974; also see
Percy and Brose 1974:18) and at the Sycamore site by Milanich (1974).
Nevertheless, the bulk of the sites are known only from meager surface
collections. The smallness of these is usually a simple result of the small
amount of material present, or visible, on the surface which, in turn, seems
to be due to two main factors: (1) the very low density of cultural refuse
apparently characteristic of most of the sites; and (2) the heavy pine needle
and leaf cover which is usually present, and which obscures to a greater or
lesser degree the few surface artifacts that exist. Adequate assessments of
the sizes and contents of sites will probably not be possible without exten-
sive excavation or other means of subsurface exploration, such as auger-
ing (Percy n. d.).
A final point which should be mentioned here is that many of the sites
are situated on lands now planted in slash pine. These lands have been
cleared by bulldozing and disc-plowed as deep as a foot below surface. Un-
doubtedly, this has caused some degree of destruction of the archaeological
sites, but without extensive excavation it is not possible to establish how
severe the disturbance has been in individual cases. Jones (1974:26-29) dis-
cusses this problem for the Upper Sweetwater area. She notes that 75-80%
of the Upper Sweetwater drainage is owned by the St. Joe Paper Company,
and that the majority of their land has been cleared and planted in slash
pine. Of the 26 sites she records in the drainage, 20 have been damaged
by slash pine planting; 2 have been severely damaged in other ways; 4 show
little disturbance. The amount of land in slash pine in this drainage shows
well in Figure 3.
In terms of surface characteristics, all sites in Figure 2 are marked
by thin scatters of artifacts, an absence of surface relief, and an absence
of midden soil, shell, and animal bone. The general classes include pot-
sherds, stone chipping debitage, and whole or broken chipped stone tools,
PERCY AND JONES
principally small stemmed or corner-notched varieties of what appear to
be projectile points (Fig. 4). There are also several occurrences of prob-
able hammerstones and one occurrence of a sandstone abrader. Six hammer-
stones are of quartzite and one of limestone (Fig. 5). The projectile points
are of flint or quartzite, principally the former. The bulk of the debitage,
which consists almost entirely of unutilized flakes, is of flint. The stone
artifacts are summarized in Tables 3 and 4.
A typological classification of the pottery from individual sites is
presented in Table 1. It is obvious that the dominant types at nearly all
of the sites are Wakulla Check Stamped and residual plain ware; although
in view of the uncontrolled nature of the samples from all sites the state-
ment that these types are dominant may be misleading. All other types,
which are Weeden Island variants of incised, punctated, complicated stamp-
ed, simple stamped, and cob-marked wares, are represented by very few
sherds and at very few sites. The same is true of the fiber-tempered
ware, Norwood Plain, which seems to be the only type not associated with
Weeden Island components.
In addition to their artifact characteristics, there are certain other
interesting features of the sites in Figure 2. They are generally rather
small. Present estimates indicate a range in size of 1 to 30 acres, but the
mean is only about 8 acres. All are situated on relatively level ground at
the edges of ridge tops and in close proximity to one or more small springs,
originating in steepheads. Generally, sites are not found in the centers of
ridges or more than several hundred feet from springs. Even where they
appear to be primarily situated alsoon themiddleor lower reaches of
streams, closer inspection also shows that they are near small tributary
The most important exception to the generalization that sites occur
on ridge tops is 8Li54, which occurs in the narrow bottomlands of Rock
Creek, just above the point where the stream flows out of the upland zone
and into the Apalachicola River Valley. It can also be seen (Fig. 2) that
one site, 8Lil9 in the Upper Sweetwater Creek drainage, is situated approx-
imately in the center of a ridge at a distance of about 800 feet from the
nearest spring. A few other sites are actually sitting on small hills, rather
than ridges, but their basic topographic situation is not significantly diff-
erent from those on the ridges.
Under natural conditions, where there is no disturbance from modern
slash pine or other farming, the ridges and hilltops are dominated by an
association of longleaf pine, scrub oak, and wire grass, while springheads,
creek bottoms, and the lower slopes of ridges and hills are characterized
by hardwood forest. Correspondingly, the soils on the crests and upper
TABLE 1. ANALYSIS OF SURFACE CERAMIC COLLECTIONS FROM UPLAND SITES IN LIBERTY AND
GADSDEN COUNTIES, APALACHICOLA DRAINAGE
T.i15 Li16 Lil7 Lil8 Lil9 Li20 Li21 Li22 Li23 Li24 Li25 Li26 Li27 Li28 Li29 Li30 Li31 Li:
32 Li33 Li34 Li35 Li36 Li37 Li38 Li39 Li40 Li44
Crooked River Comp.
Stmpd., late variety
Swift Creek Comp.
Stmpd., late variety
Thomas Simple Stmpd.
Wakulla Check Stmpd.
14 6 2 2 5 9
16 3 2 1 3
15 11 2 1 2 2
3 32 2 34 25 4 11 1 1 1
14 1 2 1 31 9 3 17 2
3 2 1
8 12 1 17
10 12 12 1
Total 59 9 16 6 6 11 5 0 1 4 2 34 2 2 5 42 2 77 2 51 1 8 38 18 1 16 8
Li45 Li46 Li47 L 8
Crooked River Comp.
Stmpd., late variety
Swift Creek Comp.
Stmpd., late variety
Thomas Simple Stmpd.
Wakulla Cneck Stmpd.
4 3 1 17 5 3 12 15
p 1 11 p p p p p 2 p
2 20 3 10 39
10 3 16 21 68 1
1 37 p p p p p 24 p
7 31 14 47 9 25 44 133 1 15 4 0 1 1 4 1 2 72 - 34 -
7 31 14
I--- ---~ ---~ ---~
- -- -7 '1 -
II_)V YI~I YI~L YI~~ Y1_~ YI~V Y~~I LI~Y YI~~ YIVV ulV~ ULV- ~-r_ r-r- --rr -~-- ---- ----- ---- --- ---- ----
Li45 Li46 Li47 Li4u
1 4 1 2 72 34 -
Li49 LA cra 2 4 a m
iL L liL.q I.~kll I.lh Ilhl Ik Ilhn I.inn I;niL I.nrs ~~oi~ ~aia ~oi ~oaa ciuai
47 9 25 44 133 1 15 4 0 1
PERCY AND JONES
S ,rrl""lI~r l"
Fig. 4. Chipped stone projectile points from upland sites (see Table 3).
SURVEY OF UPLAND LOCALES
Fig. 5. Drill, hafted scraper, knife, abrading and hammerstomes from
from upland sites (see Table 3).
INVENTORY OF SURFACE STONE ARTIFACTS
Li15 Lil6 Li7 LilS Lil9 Li20 Li21 Li22 Li23 Li24 Li25 Li26 Li27
Chips 283 5 39 34 4 3 16 13 5 21 6 B1
Cores 2 1 2
(chipped stone) 3 1 1 1
(chipped stone) 1
Total 290 6 43 34 4 3 16 13 5 22 0 7 18
Li2B Li29 Li30 Li31 Li32 Li33 Li34 L135 Li36 Li37 Li38 Li39 Li40 Li44
Chips 10 10 4 3 1 5 2 11 25 14 9 117
Cores 1 1 1 1 1
Hammerstones 1 1
(chipped stone) 1
(chipped stone) 1
(cnipped stone) 1
Total 11 10 4 3 2 0 6 2 11 26 16 0 10 121
Li45 Li46 Li47 Li48 Li49 Li5b Li51 Li52 Li53 Li54 Li56 LiS7 Li58 Li59
Chips 21 16 1 8 12 7 17 81 28 34 13 2 22
(chipped stone) 1 2 4 1
Hammerstones 7 1
Total 21 16 1 8 12 7 18 83 39 0 35 13 2 23
Li60 Li61 Li62 Li63 Li64 L1i66 Gd12 Gdl4 GdlI5 Gd6 0dl7 Gd8s8 d89
Total 1 9 2 7 33 20 -
CHARACTERISTICS OF STONE ARTIFACTS*
S H 1143 w
S4e 4J 43 -H O O
Specimen and a 4 ag -4 43 N W
Provenience O -Hq 4 1 x +g -4 N w 1 0
S0 X' 0 0 0 0 0 En 0& -H -- e C
f4) I-d .11 14 40 w W W r-4 3 0 A M
M- a, r Q En Z En h a I w
1. Chipped Stone Projectile Points (Fig. 4)
A, 8Lil5 11 13 X X X
B, 8Lil5 32 26 X X X
C, 8Lil5 13 25 X X X
D, 8Lil7 28 26 X X X
E, 8Li24 47 15 X X X
F, 8Li26 30 30 X X
G, 8Li53 47 32 X X
H, 8Li53 32 28 X X X
I, 8Li53 43 40 X X
J, 8Li53 39 35 X X X
K, 8Li59 25 13 X X X
L, 8Li64 22 33 X X X
M, 8Li66 20 21 X X X
N, 8Li51 65 33 X X
O, 8Li52 28 21 X X
P, 8Li52 29 23 X X X
2. Other Stone Artifacts (Fig. 5)
A, 8Lil7 12 6 X X
B, 8Li44 36 33 X X
C, 8Li44 48 24 X X
D, 8Li15 56 61 X X
E, 8Li38 77 57 X X
F, 8Li53 88 79 X X
G, 8Li53 100 75 X X
H, 8Li53 56 46 X X
I, 8Li53 85 51 X X
J, 8Li53 92 62 X X
K, 8Li53 106 72 X X
L, 8Li56 132 87 X X
M, 8Li32 76 35 X X
N, 8Li34 53 42 X X
0, 8Li53 65 57 X X
P, 8Lil5 81 56 X X
*Note: All points pictured in Fig. 4 have some secondary
chipping on blade edges and on both faces. Of the
artifacts pictured in Fig. 5, specimen A is a fragment
of the distal end of a drill, specimen B is a hafted
scraper, specimen C is a possible scraper or knife,
specimen D is an abrading stone, and specimens E-P are
1 9 p 2 6 32 p p p p p 20 p
SURVEY OF UPLAND LOCALES
slopes of the hills and ridges are very sandy and low in plant nutrients,
while the lower slopes and the creek bottoms and spring areas are more
silty with higher concentrations of plant nutrients. In their positions on
the edges of ridges and hilltops, the sites straddle these distinct soil and
The data obtained from survey can be supplemented by the results of
excavations at several sites, especially Torreya (8Li8) and Sycamore (8Gdl3).
Sycamore (Milanich 1974) is principally a Weeden Island house site, utilized
during the period ca. A.D. 650-1000 and the phases of Weeden Island de-
signated as 3-5 by Percy and Brose (1974:6). The latest house construction
apparently represented a Weeden Island 5 occupation, and it was radio-
carbon dated a ca. A.D. 800-1000. Unfortunately, it was not possible to
determine from the set of dates reported by Milanich just where in this
range the house most likely dated; the one building was certainly not stand-
ing for two hundred years.
Associated with the house, which was oval-shaped and approximately
30 by 20 feet, were refuse deposits indicating a variety of subsistence act-
ivities, including hunting, freshwater fishing and shellfishing, wild plant
food gathering (of hickory nuts, acorns, walnuts, and wild plums), and
corn agriculture. The house appears to represent the dwelling of a single
nuclear or small extended family. Milanich suggests it is one of many sim-
ilar, seasonal, upland habitation units, utilized from late fall to early spring
by family groups who moved to the bottomlands on the west side of the Apala-
chicola River for summer agriculture.
In addition to the Weeden Island occupation, there appear to be Nor-
wood and Deptford components at this site, although Milanich (1974:28, P1. 5)
seems to rule out the possibility of the latter, when he suggests that the
occurences of podal supports on check stamped vessels and linear check
stamped decoration are features of Northwest Florida Wakulla Check Stamped
and associated with Weeden Island. This opinion is not supported by the
stratigraphic evidence from Areas I and IV(Milanich 1974:16-17,21; Tables
4-6), where the type frequency trends, though not strongly marked, suggest
that the linear check stamped pottery more closely parallels the behavior
of the fiber-tempered pottery than that of Wakulla Check Stamped. The re-
verse should be true, if Milanich is correct. It can also be noted that, typo-
logically, the linear check stamped sherds from Sycamore as well as those
with the podal supports are within the defined ranges of characteristics of
Deptford Linear Check Stamped and Deptford Bold Check Stamped, respect-
ively (Willey 1949:354-57). We feel more conclusive evidence is needed to
demonstrate the podes from Sycamore are Weeden Island artifacts. Alter-
PERCY AND JONES
natively, it is equally logical that they are not Weeden Island traits, but
markers of a possible Deptford component.
The Torreya site in Liberty County is a Weeden Island habitation site
with at least 13 distinct house or refuse-depositing areas, each approx-
imately the same size as the one at Sycamore. The same range of sub-
sistence activities is represented, except that cultigens have not been found.
A few undecorated fiber-tempered sherds have been recovered, suggesting
a Norwood component, but there is no evidence of Deptford; neither the
linear check stamping nor the podal supports are present in the collections
gathered to date. The Weeden Island occupation seems to begin in middle
Weeden Island (Weeden Island 3 is tentatively suggested as the earliest
phase), as at Sycamore, with five or six house units. The site is then
abandoned and re-occupied in a spatially separate area in very late Weeden
Island (Weeden Island 5); the number of house units again seems to be five
or six. At the end of Weeden Island 5, the site is again abandoned and is
not subsequently re-occupied in Fort Walton. The overall picture sketched
here may change as more of the collection from the 1972 field season is
analyzed. In particular, the question of the abandonment between phases 3
and 5 needs further study, and present statements on the point are best
considered as hypotheses. The analysis of the materials from the 1971 field
season, which represent the phase 5 occupation, has been completed, and
the totals for the different ceramic types in this collection are presented in
Associated with 8Li8 is a small Weeden Island burial mound (8Li5,
Nature Trail Mound). The Weeden Island classification is based on materials
excavated by C. B. Moore in 1918 (Moore 1918:554-55, Pl. XV). Moore
called the site "Mound near Rock Bluff Landing. He discovered thirteen
badly decayed burials and an east side pottery cache. The latter included
"vessels and parts of vessels bearing the check stamp, the complicated
stamp, several with interesting incised designs..." (555), and one exotic
double-globed vessel with cut-outs and a combination of incised and im-
ressed or pinched decoration (P1. XV).
Moore also conducted excavations in one of the mounds at the Aspa-
laga site (8Gdl), which he described as consisting of three sand mounds
and a shell and black-earth midden (Moore 1903:481-88). On the basis of
Moore's descriptions and illustrations, Willey (1949:257-58) classified the
site as Santa Rosa-Swift Creek, overlapping into Weeden Island I; however,
he avoided a specific classification of the mound excavated by Moore, which
was a rather elaborate burial mound.
More recently, William Sears discovered a fourth mound at Aspalaga;
he also made a contour map of the entire site and surface collected in the
SURVEY OF UPLAND LOCALES
Table 5. Frequency of ceramic types from the north end of the Torreya site
(8Li8), 1971 field season, including both surface and excavated materials
Wakulla Check Stamped 3
Residual Plain 1
Swift Creek Complicated
Stamped, Late Variety
Old Bay Complicated Compl. St'd
Weeden Island Incised
Weeden Island Punctated
*643 Tucker Ridge Pinched 1
767 Fort Walton Incised 1
Single horizontal-line incised 17
15 Multiple parallel-line incised 5
5 Unidentified incised 5
1 Unidentified punctated 2
12 Brushed 2
3 unidentified decorated 12
1 indeterminate sand-tempered 570
8 Norwood Plain 3
village area. He is of the opinion that the major use of the site occurred
from Early Swift Creek through early Weeden Island (personal communicat-
ion from William H. Sears, Department of Anthropology, Florida Atlantic
University, Boca Raton, February 11, 1975). He notes that Santa Rosa cer-
amics are absent in his village collection, and that Early Swift Creek cer-
amics are as heavily represented as Weeden Island types. Late Swif Creek
types are also present, and Sears feels that Phelps' Early and Late Swift
Creek phases (Phelps 1969) more adequately characterize the ceramic se-
quence at Aspalaga than Willey's single Santa Rosa-Swift Creek category.
The details of Sears' work will appear in a forthcoming publication by him.
Much work remains to be done before prehistoric Indian occupations
of the upland zone in western Gadsden and Liberty counties can be reliably
reconstructed. Large tracts of this area are still unsurveyed, and very few
of the presently known sites have been more than cursorily examined. Only
one, Sycamore, has been excavated intensively enough to produce a repre-
sentative sample of features and artifacts. Thus, although a few impressions
and hypotheses can be offered, it would be premature to expect that definite
conclusions about the use of the uplands could be drawn from present data.
One point that is worthwhile mentioning is that site density is much
higher than originally suspected. The density is apparently variable from one
locale to another--for example, compare the upper Rock Creek drainage with
the Upper Sweetwater Creek drainage--but on present data it averages out
at four per square mile. The variability may be partly a function of the sur-
vey coverage. In the example cited above, all of Upper Sweetwater Creek has
been surveyed, but a significant proportion of upper Rock Creek has not.
PERCY AND JONES
When the Florida State University research was begun in 1971, it was thought
that the uplands might be a marginal zone, at least for late prehistoric agri-
cultural groups with very few traces of settlement. This impression has been
Practically all of the sites which have been discovered so far repre-
sent Weeden Island occupations. No preceramic Archaic materials have been
recovered anywhere. Norwood fiber-tempered pottery has been found at a
number of places, and Norwood components appear to be the second most
frequent in the uplands. Deptford, Early Swift Creek and Late Swift Creek
occupations are represented by only one component each. The last two both
seem to be present at Aspalaga, and a Deptford occupation may exist at
Sycamore. One or two Fort Walton sherds have been found at the Torreya
site, but there is no other sign of a Fort Walton occupation, and no Fort
Walton materials have come from any of the 57 other sites.
In the collections from the Weeden Island components, the outstand-
ing fact is the predominance of residual plain and Wakulla Check Stamped
pottery and the low incidence of other Weeden Island types. All of the com-
ponents, with the possible exception of that at Aspalaga which is as yet un-
explored, seem to fall in the middle to late Weeden Island range--in the
range of phases 3-5 suggested by Percy and Brose (1974:6)--and mostly
towards the late end (phase 5). The radiocarbon dates from Sycamore sug-
est that the time period for these components would be ca. A.D. 650-1000.
The Weeden Island 5 ceramic complex is well represented at the Torreya
site (northern end) and at Sycamore, where it is associated with the struct-
ure recovered by Milanich.
With two principal exceptions (8Lil9 and 8Li54) the Weeden Island sites
tend to occur along the edges of ridge tops, near to springs and at the in-
terfaces of distinct soil and vegetation zones. In these positions, they are
close to patches of well-watered, silty or loamy soil, existing around the
springs, and still situated on relatively level, well-drained surfaces that
are convenient for settlements. This is the situation today, and there is
no evidence that the local environment has changed significantly since Weed-
Percy and Brose (1974:19-21) have hypothesized that the sudden appear-
ance of a large number of sites in the uplands east of the Apalachicola Riv-
er, during middle and late Weeden Island, reflects a significantly greater
dependence upon agriculture and a dramatically increased need for agricult-
ural land at this time. The principal crops being grown were probably maize
and squash; possibly also beans, but the last may not have appeared any-
where in the Southeast United States until after A.D. 1000 (Yarnell 1964:118-
SURVEY OF UPLAND LOCALES
19). Percy and Brose argue that the agricultural system was one of shift-
ing cultivation by small family groups, who were required to alter the lo-
cations of their settlements after periods of a few years because of the ex-
haustion of local soils. This pattern could explain the hypothetical abandon-
ment and re-occupation of the Torreya site. When in use, settlements were
occupied year-round, and populations did not move seasonally to different
base camps, as Milanich has suggested (1974:34-35). Specifically, there is
no evidence that groups moved down from the uplands to the Apalachicola
River bottomlands for summer farming. Upland sites, according to Percy
and Brose' s hypothesis, were as much agricultural settlements as the bottom-
land ones, and the need for land was so great that essentially all environ-
mental zones in the upper Apalachicola basin were used for farming.
The occupants of upland locales probably attempted to concentrate their
farming in the small bottomlands immediately in front and alongside of spring-
heads, where soil nutrients were not as quickly exhausted as in the sandy
pineland soils on the ridge tops. This would explain the tendency for Weeden
Island sites to concentrate where they do. Throughout the region of heav-
ily dissected uplands in northwestern Liberty and southwestern Gadsden
counties, it is the areas immediately around springheads that are natural-
ly best suited for agriculture. The creek bottoms downstream from the
springs have similar soils, but are not as suitable because they are too nar-
row. Except in their lowermost reaches, creeks are deeply entrenched with
high, steep banks that permit very little overflow and deposition of new soil;
their steepness also means that erosion is a problem, and this prevents the
build-up of thick, nutrient-rich, humic zones. Around springheads, there is
usually some land which is less steeply sloped and, therefore, less subject
to erosion, as well as physically more convenient for gardening activity.
However, even these are small areas, and the average size patch associ-
ated with one springhead would probably only be enough for one family's
garden. On this assumption, sites with several families in residence at one
time would be those in close proximity to more than one spring.
The foregoing statements are all hypothetical and obviously must be
tested by excavations.. It is suggested that all of these sites are agricult-
ural base camps, varying in size from one house unit, like Sycamore, to
several house units, like Torreya. Not all would have been in use at the
same time. Rather, some would have been unoccupied, while the associated
agricultural lands were lying fallow. This hypothesis, as it relates to the
Torreya site, supplants Percy's earlier proposal that Torreya was occupied
on a seasonal basis for hunting and gathering activities (Percy 1972c:ZZ).
Judging from the results of excavation at Torreya and Sycamore, the
Weeden Island agricultural economy, even at the very end of Weeden Island,
PERCY AND JONES
was supplemented by hunting, fishing, shellfishing, and wild plant food gath-
ering (Milanich 1974:31-34; Percy 1974:68,81). The products of these activit-
ies were not necessarily of secondary importance to cultivated plants in the
daily diet, but, on hypothesis, they could not by themselves have supported
the later much larger Weeden Island populations.
Scientifically designed analyses of the lithic materials from the upland
sites have not been attempted. The most detailed functional classification of
lithics is Milanich's (1974:22-26) for Sycamore, but even this is almost to-
tally hypothetical, since most of the assessments of function are not explained.
The sources of flint may have been residual limestone boulders con-
tained in Miocene deposits covering the uplands and exposed through erosion
of overlying Plio-Pleistocene and Recent sediments; or flint may have been
obtained from limestone boulders or exposures of intact limestone format-
ions, west of the Apalachicola in northern Jackson and Washington counties
and eastern Holmes County. In either case, it appears, both from excavated
and surface collections, that the demand for flint was not great at the up-
land Weeden Island sites. The density of tools and debitage is consistently
light at practically all sites, although one exception is 8Lil5, which pro-
duced a large number of chips, as well as a relatively large number of
sherds. Either this site was more intensively occupied than most of the
others, or, perhaps, it was established near a residual limestone boulder
containing flint and served principally as a quarry site. Other rocks, such
as quartz, quartzite, sandstone, and granite, were utilized in addition to
flint. All were obtainable locally in the uplands and probably were picked
up in creek beds where they had eroded out of Miocene deposits.
Very few burial mounds have been located in the uplands, and the
only one known definitely to be a Weeden Island mound is 8Li5, the Nature
Trail Mound, associated with the Torreya site. The significance of this
lack of mounds is not clear. It could be a point in favor of a seasonal occ-
upation hypothesis. On the other hand, it may simply be that the mounds
serving the area have not yet been found. For example, there are reports
of mounds in the lower portion of the Sweetwater drainage, which is still
So little data has been accumulated on Norwood, Deptford, and Swift
Creek components that there is practically nothing that can be said about
the use of the uplands by these cultures. The Aspalaga site promises to
hold significant data on Swift Creek and is also the only site discovered to
date that seems likely to produce information on an early Weeden Island
The work reported in this paper represents only the barest beginnings
SURVEY OF UPLAND LOCALES
ot a study of the upland zone in the upper Apalachicola basin, and many
more years of work lie ahead just to answer some of the questions raised
so far. In general, future studies should be conducted along three principal
lines: (1) continuation of the survey of southwestern Gadsden and northwest-
ern Gadsden and northwestern Liberty counties, until the entire upland area
within the Apalachicola drainage is completed; (2) intensive investigation of
individual sites in this same area; and (3) comparative studies of other up-
land regions in Panhandle Florida, as well as comparative studies of other
environmental zones in the upper Apalachicola basin. Hopefully, future work
will also be designed to test and refine some of the hypotheses generated
from the work already accomplished.
Jones, Mary Katherine
1974 Archaeological survey and excavation in the upper Sweetwater
Creek drainage of Liberty County, Florida. Unpublished Master's
thesis, Florida State University.
Milanich, Jerald T.
1974 Life in a 9th century Indian household. A Weeden Island fall-
winter site on the upper Apalachicola River, Florida. Bureau
of Historic Sites and Properties, Bulletin no. 4, 1-44. Florida
Department of State. Tallahassee.
Moore, Clarence B.
1903 Certain aboriginal mounds of the Apalachicola River. Journal of
the Academy of Natural Sciences of Philadelphia, second series,
vol. 12, 439-92. Philadelphia.
1918 The northwestern Florida coast revisited. Journal of the Acad-
emy of Natural Sciences of Philadelphia, second series, vol. 16,
Percy, George W.
1971a Current research: Florida. Newsletter of the Southeastern Arch-
aeological Conference, vol. 15, no. 1, 7-8. Morgantown, W. Va.
1971b Preliminary report to the Division of Recreation and Parks,
Department of Natural Resources, State of Florida, on archaeolog-
ical work in Torreya State Park during the year 1971 by the Depart-
ment of Anthropology at Florida State University. Tallahassee.
1972a Current Research: Florida. Newsletter of the Southeastern Arch-
aeological Conference, vol. 16, no. 1, 3-6. Morgantown, W. Va.
PERCY AND JONES
1972b Current Research: Florida. Newsletter of the Southeastern Arch-
aeological Conference, vol. 16, no. 2, 9-10. Morgantown, W. Va.
1972c A preliminary report on recent archaeological investigations in
Torreya State Park, Liberty County, Florida. Paper presented
at the 24th annual meeting of the Florida Anthropological Society,
Winter Park, Florida. (Mimeographed).
1974 A review of evidence for prehistoric Indian use of animals in
Northwest Florida. Bureau of Historic Sites and Properties,
Bulletin no. 4, 65-93. Florida Department of State. Tallahassee.
1976 Use of a mechanical earth auger at the Torreya site, Liberty
County, Florida. Florida Anthropologist, vol. 29, no. 1, pp. 24-32.
Percy, George W., and David S. Brose
1974 Weeden Island ecology, subsistence, and village life in Northwest
Florida. Paper presented at the 39th Annual Meeting of the
Society for American Archaeology, May 2-4, Washington, D.C.
Phelps, David S.
1969 Swift Creek and Santa Rosa in Northwest Florida. The Institute
of Archaeology and Anthropology, Notebook, vol. 1, nos. 6-9,
14-24. Columbia, South Carolina.
Puri, Harbans S. and Robert O. Vernon
1964 Summary of the geology of Florida and a guidebook to the classic
exposures. Florida Geological Survey, Special Publication No.
5 (revised). Tallahassee.
1967 A review of techniques for archaeological sampling. In A Guide
to Field Methods in Archaeology, ed. by Heizer, Robert F. and
John A. Graham, 181-97. The National Press. Palo Alto.
SPANISH TRADE PIPES FROM MARION COUNTY, FLORIDA
Wilfred T. Neill and George R. Ferguson
In the Sixteenth Century, the practice to tobacco-smoking was carried
from the New World to Europe, where for a time it took two different
courses. The English soon became pipe-smokers, but not so the Spanish,
who preferred cigars. This European dichotomy probably reflected the cir-
cumstance that the English learned smoking from the pipe-using Indians of
the eastern United States, the Spanish from Antillean tribes who rolled their
tobacco in a leaf. Whether through their own preference or that of their abor-
iginal customers, the English brought many trade pipes to the New World,
the Spanish comparatively few. Considerable attention has been given in the
Florida literature to English and Early American trade pipes, but not to
Spanish. It is therefore of interest to describe three tobacco pipes, recovered
from sites in Marion County, Florida, under conditions suggesting Spanish
Brinton (1859:172) remarked, "On the opposite banks of Silver Springs
run, respectively a quarter mile and a mile and a half below the head, there
are two tumuli. Pottery, axes, and arrow-heads abound in the vicinity... "
The latter of these two mounds, on the north side of the stream, was hauled
away for its sand content, probably in the 1930s. The remnant of this mound,
a low sandy knoll, produced a pipe (Fig. 1, a), a few spalls of pink or white
flint, 12 sherds of St. Johns Plain, and two sherds of St. Johns Check Stamped.
The pipe, made of fired clay, is clearly of non-Indian origin. The brick-
like color and the hardness of the material suggest a firing process superior
to the aboriginal one, and a faint trace of a longitudinal seam reveals the ob-
ject to have been mould-made. The shape is the so-called "elbow". Bowl and
stem are ribbed. Artifact typology does not negate the supposition that the
pipe and the sherds were coeval, for St. Johns Plain and Check Stamped pot-
tery continued into historic times.
A second pipe (Fig. 1, b), very similar to the first, was found in a
sand mound located about two miles northeast of Lynne. The site may be
Mr-43 in the list of Goggin (1952:96). During the 1930s, much of this mound
was removed by a road that cut through it. The remnant has been plentifully
potholed. The upper sand of this remnant yielded the aforesaid pipe, a few
flint spalls, bits of charcoal, one small "thumbnail" scraper of local flint,
nine sherds of St. Johns Plain, five of St. Johns Check Stamped, four of
Pasco Plain, and four of a smooth, well-made, sand-tempered plain.
A third pipe, closely similar to the two figured herein, was recovered
by a diver from the bottom of Silver Springs, at the head of the springs.
Florida Anthropologist, vol. 29, no. 3, September 1976
NEILL .AND FERGUSON
This third specimen has been figured by Martin (1966:47). It was accomp-
anied by a large number of St. Johns Plain sherds, many of them red-paint-
ed. Of course, contemporaneity of the pipe and the sherds cannot be dem-
There seems to be no literature reference to Florida pipes precisely
like the Marion County ones, but Willey (1949:PI. 57, v-x) figured three
generically similar specimens from the Safety Harbor site, Pinellas County.
These three dated from the Safety Harbor Period. Willey categorized them
simply as European, but a Spanish origin is likely.
Seminole sites are numerous in and near Marion County (Neill 1956:
ZZ-24), and without exception they have yielded fragments of long-stemmed
"church-warden" pipes. These are of white ball-clay (commonly but inacc-
urately called kaolin in the Florida Literature). Only one Seminole site in
this region has yielded an "elbow" pipe of non-Indian origin: Osceola's
Village in Marion County (Neill 1955:243, Fig. E). This site had a pre-Sem-
inole component, some artifacts of which had been brought to the surface by
repeated plowing and commingled there with Seminole debris. Thus, the pipe
from Osceola's Village site may be of pre-Seminole origin. It differs from
the three presumably Spanish pipes in being thicker and cruder, glazed, and
without longitudinal ribbing or other ornamentation.
Our three presumably Spanish trade pipes are from sites in the North-
ern St. Johns Archeological Area. Within that area, early historic aboriginal
pipes are of the "elbow" type, and most of them have been recovered from
burial mounds (Goggin 1952:128). It is therefore interesting to note that our
specimens are "elbows", and two of them are from mounds.
In conclusion, we regard these trade pipes to date from the St Johns
IIc period, and to have been of Spanish origin.
Brinton, Daniel G.
1859. Notes on the Floridian Peninsula, Its Literary History, Indian
Tribes and Antiquities. Philadelphia. Joseph Sabin, publisher.
Goggin, John M.
1952. Space and Time Perspective in Northern St. Johns Archeology,
Florida. Yale University Publications in Anthropology, No. 47.
Martin, Richard A.
1966. Eternal Spring; Man's 10,000 Years of History at Florida's
Silver Springs. Great Outdoors. St Petersburg, Florida.
Neill, Wilfred T.
1955. The Site of Osceola's Village in Marion County, Florida. The
Florida Historical Quarterly, Vol. 33, Nos. 3-4, pp. 240-246.
1956. Florida's Seminole Indians. Ross Allen's Reptile Institute,
Willey, Gordon R.
1949. Archeology of the Florida Gulf Coast. Smithsonian Miscellaneous
Collections, Vol. 113. Washington, D.C.
~.L42L .f- I 2 3 _
Fig. 1. Two trade pipes from Marion County, Florida
May 15, 1975
THE SEMINOLE PUMPKIN
Wilfred T. Neill
The so-called Seminole pumpkin merits the attention of ethnobotanists.
Strictly speaking, it is not a pumpkin but a cushaw, Cucurbita moschata. Un-
der cultivation by American Indians, this species diverged into several distinc-
tive varieties, large or small, smooth-shelled or warty, uniform yellow or
green-and-white striped, cold-hardy or frost-sensitive.
The species, although not necessarily the "Seminole pumpkin" variety,
may first have been domesticated in Mexico or Central America (Sauer 1950:
504-505). The earliest Mesoamerican occurrence is about 6000 B.C., in the
El Riego phase of Puebla, Mexico (Willey 1966:81-84). It is also an old culti-
gen on the Peruvian coast, where it has been recovered from a site of the Pre-
ceramic Period V, a period dated 4200-2500 B.C. (Willey 1971:93). In the
Peruvian Andes, it is known not only from archeologically recovered specimens
but also from ceramic portrayals of the Moche (equal Mochica) civilization,
The species has been reported ethnographically from tropical South Amer-
ica, Mexico, and the Southwestern United States (Meighan et al. 1958:138).
Beverley (1705:27) also described its cultivation by Algonquian Indians of Vir-
ginia. The common name of cushaw is in fact from the Algonquian. Moche ves-
sels portray a crook-necked cushaw and most ethnographic reports, including
those from the United States, also relate to a crook-necked variety. The Sem-
inole pumpkin, however, usually has a globular fruit. This is golden in color,
not green-and-white striped like the cushaws of the Algonquians. The plant is
remarkable in growing as a vine that climbs large trees, to dangle the fruits
perhaps 40 feet above the ground.
In spite of the common name, the Seminole pumpkin probably was unknown
to the Seminole Indians prior to their arrival in Florida; for it is a subtropical
variety, unsuited to the climate of the Georgia-Alabama region whence came the
ancestors of the Seminole. Presumably, then, this was a cultigen of the Tim-
ucua, who preceded the Seminole in Florida. The paintings of Jacques Le Moyne,
made in 1564, show some kind of cucurbit in use by the Timucua as a container
(Lorant, ed. 1946:32, 75), but it has a very large, flattened-globular fruit and
is not a Seminole pumpkin. Of course, Le Moyne was only in northeast Florida,
probably too far north for this plant to be grown. There are a few early ref-
erences to pumpkins among the Timucua and Apalachee, but they probably re-
late to some variety of the true pumpkin, Cucurbita pepo. Accordingly, an ethno-
botanical account of the Seminole pumpkin must begin in Seminole times.
Probably in the early 1800s, Seminole Indians built a village on a large,
Florida Anthropologist, vol. 29, no. 3, September 1976
spring-fed river of Citrus County, Florida. They gave it a name that today is
rendered Chassahowitzka. Read (1934:4) translated this as "hanging pump-
kins" in Creek, the language of almost all Seminole toponyms in Central Flor-
ida. The name first appeared on a map in 1839, at which time it designated
the aforesaid river. The application of the name has since been extended to a
large swamp just south of the stream, and to a point of land near the stream's
As I have pointed out in greater detail elsewhere (Neill, in press) the
local vernacular pronunciation of a Seminole toponym, as retained by rural
whites of limited literacy, is likely to give a better idea of the original toponym
than do the various spellings encountered on old maps and documents. In this
connection, it is interesting to note that the local vernacular pronunciation of
Chassahowitzka, "Chassy-wisky", is virtually identical with the Creek-Semi-
nole name for the cushaw, a name that does indeed have the literal meaning
of "hanging pumpkins".
Not surprisingly, there are no ethnobotanical references to the cushaw
during the Seminole Wars, nor during the times shortly thereafter, when the
Indians, still hostile, had withdrawn to the South Florida wilderness. However,
at a later date the Seminoles recalled the way American troops would amuse
themselves in wartime by shooting at the cushaw stems in order to bring the
fruits down (Moore-Willson 1896:82).
MacCauley (1887:504) mentioned pumpkins among the Seminole. Prob-
ably his reference was to the cushaw, although the Indians also grew the true
pumpkin. Small (1933:1287) remarked that the Seminole pumpkin grew around
both occupied and abandoned Indian campsites. It is uncertain how long the
species could survive in the wild after cultivation of it had ceased at some In-
dian or pioneer white settlement. Since the plant is a cultigen, yet does grow
unattended at a few scattered localities in Central Florida, perhaps its pres-
ence there indicated the existence of interesting sites nearby. Certain it is
that another cultigen, the sour orange, has persisted from Seminole and pio-
neer times in Central Florida.
Gifford (1944) considered the cushaw to have been very important to the
Seminole in South Florida, as well as to early white settlers in the same area.
He noted that the Indians would plant the seeds at the base of a dead tree. The
vine would, of course, receive far more light, and thus grow more luxuriantly,
on a dead tree than on a living one. Sometimes a Seminole would girdle a liv-
ing tree to kill it, and so render it more suitable for growing of cushaws. The
Seminolepumpkinis still raised by the Indians in South Florida. Green or ripe,
the fruit may be boiled or roasted. Alternatively, the flesh may be cut into
strips and dried over coals. The flesh is palatable, rich in starch and sugar;
and the plant may be grown on trees at the edges of (or within) the Seminole
cornfields. No wonder that up until quite recent times, the Indians prized
this unusual variety of cushaw.
The foregoing account of the Seminole pumpkin has of necessity been
brief. The variety has not been recovered archaeologically. The cultivation
of it by the historic Timucua is hard to analyze, for the early writers onFlor-
ida Indian agriculture did not carefully distinguish cushaws from pumpkins
and squashes, sometimes not even from gourds. It may not be possible to
compare the Seminole pumpkin with other United States aboriginal varieties,
for the latter probably have disappeared, replaced by a dozen or more vari-
eties produced by modern horticulturists. Perhaps some useful information
would emerge from a morphological and genetic comparison of the Florida
variety with cushaws still grown in remoter parts of Mesoamerica and South
America. In this connection, Bukasov ( 1930) made a start toward an evo-
lutionarily meaningful breakdown of Cucurbita moschata into races.
1705 The History and Present State of Virginia, in Four Parts ... By
a Native and Inhabitant of the Place. Part Two. London.
The Cultivated Plants of Mexico, Guatemala, and Colombia. Bul-
letin of Applied Botany and Plant Breeding, Supplement 57.
Gifford, J. C.
1944 Five Plants Essential to the Indians and Early Settlers of Florida.
Tequesta, No. 4, pp. 36-44. Coral Gables.
Lorant, Stefan (editor)
1946 The New World. Duell, Sloan and Pearce, New York.
The Seminoles of Florida. Annual Report, Bureau of American
Ethnology, No. 5, pp. 475-553. Washington.
SW., D. M. Pendergast, B. K. Swartz, Jr., and M. D. Wissler
Ecological Interpretation in Archaeology: Part II. American
Antiquity, Vol. XX, No. 2, pp. 131-150. Salt Lake City.
1896 The Seminoles of Florida. New York.
Neill, Wilfred T.
In press Archeology and a Science of Man. Columbia University
Press. New York and London.
Read, William A.
1934 Florida Place-Names of Indian Origin and Seminole Personal
Names. Louisiana State University Press. Baton Rouge.
Sauer, Carl O.
1950 Cultivated Plants of South and Central America. In J. H.
Stewart, ed., Handbook of South American Indians, Vol. 6.
Bureau of American Ethnology Bulletin 143, pp. 487-543.
Small, John K.
1933 Manual of Southeastern Flora. Published by the Author, New
Willey, Gordon R.
1966 An Introduction to American Archaeology. Vol. 1. North and
Middle America. Prentice-Hall, Inc. Englewood Cliffs, New
1971 An Introduction to American Archaeology. Vol. 2. South Amer-
ica. Prentice-Hall, Inc. Englewood Cliffs, New jersey.
New Port Richey, Florida
May 28, 1975
Back issues of the FLORIDA ANTHROPOLOGIST, if available, may
be purchased from the Treasurer at $6.00 per volume ($1.50 for single and
$3.00 for double numbers. Vols. 1; 7:2-3; 8:1-2, 4; 9; 10; 11:1-2; 13:1;
17:1; 18:1; 21:1, 4; 24:1; 25:3; and 26:1 are out of print.
Copies of the FLORIDA ANTHROPOLOGICAL SOCIETY PUBLICATIONS,
a monograph series, may be purchased as follows: Numbers 5, 6, and 7 at
$2. 00 each; Number 8 at $3. 00 each. Others are out of print.
Orders with remittances for any of the above should be sent to the
Treasurer whose address is given on the inside of the front cover. The
Society has no resale discount for dealers or resale agents.
INFORMATION FOR AUTHORS
THE FLORIDA ANTHROPOLOGIST publishes original papers in all
subfields of anthropology with an emphasis on archaeology. Contributions
from allied disciplines are acceptable when concerned with anthropological
problems. The journal's geographical scope, in order of importance, in-
cludes Florida, Southeastern United States, Caribbean, and related regions.
Manuscripts should be double spaced and typed on one side only of an
8 1/2 by 11 inch sheet leaving at least a 1 inch margin on all sides. Contri-
butors are referred to American Antiquity and the Chicago Manual of Style
for matters of style and reference. Textual references should be similar to
"Smith 1964: 78-80; Jones 1969: Pl. II, b). No footnotes. In the "Refer-
ence Cited" section, at the end of an article, do not abbreviate names of ser-
ies or journals. Books need full title, publisher's name, and place.
All illustrative material-line drawings (in India ink) and photographs-
should be included in one numbered series of "Figures" suitable for a maxi-
mum reduction to 6 1/2 by 8 1/2 inches. Parts of figures are designated by
lower case letters. Figures must be accompanied by brief descriptive cap-
tions. All tabular material must be included in separately numbered "Tables. "
Twenty-five reprints without covers of each article are furnished free
to each author or group of authors.
FLORIDA ANTHROPOLOGICAL SOCIETY CHAPTER DIRECTORS
Northwest Florida Chapter
Thomas C. Watson
203 Carolyn Avenue
Panama City, FL 32401
Northeast Florida Chapter
Jerry J. Hyde
4233 Oristano Road
Jacksonville, FL 32210
Central Florida Chapter
Arthur F. Dreves
2414 Summerfield Road
Winter Park, FL 32789
Tampa Bay Chapter
1020-4th Street North
St. Petersburg, FL 33701
Indian River Chapter
Cliff E. Mattox
P.O. Box 521
Cocoa Beach, FL 32931
Broward County Chapter
Joan E. Lynn
5430 S.W. 36th Street
W. Hollywood, FL 33023
Palm Beach County Chapter
Joel M. Baskin
204 N.W. 1st Avenue
Boynton Beach, FL 33435
Dade County Chapter
Fredrick T. Huna
3031 N.W. 171st Street
Opa Locka, FL 33055